Eyebrow ($BROW) is an AI agent security.

AI agents install third-party skills, MCP servers, plugins, hooks... that run with full access to credentials, wallets, code, etc. There is usually no lockfile, no ongoing review & no detection of post-install changes (sleeper attacks). Eyebrow is the integrity layer: it checks everything installed across tools, fingerprints it into a commit-able lockfile, detects drift the moment an approved artifact changes, provides runtime MCP firewall, sandboxing, and audit logs.

It gives a public “eyebrow badge” for publishers to prove their artifact is unchanged since verification.

As far as the token goes:

-Hold $BROW for continuous monitoring / always-on drift detection.

-Publisher bond for live “Verified by eyebrow” badges (lock $BROW as collateral; longer lock = lower requirement; slash + burn if malicious code is shipped under a green badge

Future: metered oracle fees for agent trust lookups, marketplace fees, buyback-and-burns.

Momentum: it found a possible vulnerability in the Venice AI skill repo 3 days ago (https://x.com/eyebrowCC/status/2099485049057878510), Vlad tweeting about AI doing cybersecurity (https://x.com/vladtenev/status/2100412583253819705)